Pushing the limits Regulator

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm trying to get the voltage regulator to working from Peter Daniel. For some reason the BSP129 Current source does not work. The BSP129 are new but give zero current.

Also tried to connect the gate and source together with 15v at the drain still same problem.

Pin 1 is gate
Pin 2 is drain
Pin 3 is source
 

Attachments

  • Reg.jpg
    Reg.jpg
    155.6 KB · Views: 2,371
Member
Joined 2003
Paid Member
Extending the Bobken Regulator...

This is really a question for Bobken & for Peter D., but I'm sure others may wish to chime in.

I'd like to try the 'Bobken' regulator in a phono stage. It currently uses a not-to-dissimilar shunt regulator and starts with 22v and regulates it down to 15v.

Will this work ok as long as I don't overheat/overload the two main active devices, the BSP129 & the ZVN3310? I need to measure the current draw of the existing circuit & if it looks excessive for the devices, I may go to one regulator per active device per channel (2 active devices per channel, 4 +- regulators total) to keep within the device limits.

This of course also begs the question about the negative rail regulators. The ZVN3310 has a compliment, the ZVP3310. Can I create a negative rail version of this regulator by:

1. Reversing the orientation of the current source (BSP129 + current set resistor)?

2. Use the ZVP3310 instead of the ZVN3310 as the shunt device?

3. Reorient the power to the AD811 so that the negative supply pin is to the rail and the positive to ground? (of course, the orientation of the NP BG .01 NZ would stay the same with respect to the positive and negative supply pins.)

4. Reverse the orientation of the LM285-2.5v and of course, the polarized input capacitor and the NP output one (as NP BG's need to be oriented a certain way for the best sound)?

Bob, Peter, others, will this work ok? Has anyone else tried this?

TIA!

Greg in Mississippi
 
Peter,

I do not recall asking you about why you are calling this shunt regulator topology the “Bobken regulator”.

There does not have to be a complementary P channel for replacing the BSP129. The shunt regulator can be broken down into two key sections, both of which behave like two terminal devices. There is a current source and a shunt regulator, which behaves like a synthesized zener diode but with much better operating parameters. As long as the correct polarity is maintained for the circuit elements the current source can be moved to the negative rail with the BSP129 drain connected to the zener anode terminal and the zener cathode terminal is connected to ground thus providing a negative rail version of the regulator.
 

Attachments

  • SHUNTREG (Schematic).pdf
    2.6 KB · Views: 791
Last edited:
Peter,

That post was directed at Bobken. I am assuming that you were not given the full history of the design.

Changing resistor types, capacitor types and the output mosfet to another mosfet with a different part number does not constitute a new original design, or a progression from the original design topology. I used several types of mosfet with this design including the Zetex part. I often fit bulk foil (lately z-foil) resistors to these regulator designs when requested, and regularly used to fit Black Gate capacitors when they were readily available.
 
Hello samoloko,

The BSP129 is a depletion mosfet and cannot be replaced with an enhancement mosfet without additional bias circuitry.

The voltage across the CCS resistor will vary depending on the set current and the operating temperature of the BSS129. There is also batch variability to consider. It would typically be in the region of 0.25v to 0.35v when set up to pass 100ma.

With a typical regulator set up the AD811 bandwidth spec will be around 100MHz but this is where the error amp ceases to regulate and it does not display the same regulation performance all the way up to this frequency. Still its regulation capability is around 10 to 20 times better than typical three terminal regulators and the transient response and settling time are both much faster.

Regards
Paul
 
Last edited:
Hello samoloko,

Consider the regulator as two distinct functional circuits.

The first section is the current source, in this instance the BSP129. This can be replaced with other current source circuits and there are a number of different circuits posted on the audio forums. Also Walt Jung did several articles on current source design that you could refer to. Sorry I do not have the links.

I have tried several current source topologies with this regulator topology successfully, but I should warn you that you are dealing with a wide bandwidth regulator circuit, and it is possible that some current source circuit topologies could induce instability in the regulator. With this in mind I would suggest that you check your revised design with an oscilloscope to ensure stability under all the load conditions you will subject the regulator to.

The second section is the regulator section and this should remain as shown as it is the basis of the high performance of the regulator circuit.

Regards
Paul
 
Hello samoloko,

The LM317 alone is not a good choice for a wideband current source, as the output impedance will reduce with rising frequency above approx 6KHz. This fall off of impedance with rising frequency is caused by the falling loop gain of the LM317 error amplifier. The supply line rejection is dependant on the current source maintaining a high impedance, compared to the impedance of the regulator section. As the current source impedance drops so does the supply line rejection. This means that the supply line rejection gets worse as you move into the RF spectrum. The sonic signature of the LM317 is still apparent. Depletion mode mosfets also display a drop in impedance at high frequencies but not as abruptly as the LM317 and I prefer the depletion mosfet sonic signature.

There are better current source alternatives but you will have to do a little trial work yourself. Sorry I cannot advise you about the ones I use, as they are the subject of non-disclosure agreements with some of my consultancy clients, and I do not wish to be sued for breach of contract.

I originally designed this topology configuration specifically for the AD811 because I wanted better regulation over a wider bandwidth than alternative options available at the time. Since this design’s inception I have only worked on increasing the bandwidth, reducing the regulator noise and improving the transient response and settling time, so I cannot advise from personal experience on using slower opamps with less bandwidth. Reducing the bandwidth and speed of the regulator by fitting a lesser opamp may or may not work well with your chosen current source, however it is going to reduce the performance at high frequencies which sort of negates the reason for designing the circuit in the first place. There are too many variables to factor into the equation to make recommendations without trial, which is why I recommended using an oscilloscope to check for stability.

Regards
Paul
 
Last edited:
maximus

I am very curious to try your ad811 regulator
Is there a place of more information about It or I can use post1
what Is V out range and current range ; what application Is best from supply analog line circuits, supply analog part of DAC chips, supply digital part of DAC chips
 
Hello samoloko,

I have not published applications information about this regulator topology, as it was a commercial product until I re-designed it to increase speed and operating bandwidth, as well as reducing noise. The circuit was published by Peter Daniels after forum member Bobken sent him the circuit information to try with his TDA1543 Dac project. Peter assumed it was Bobken’s circuit but this was erroneous as I supplied Bobken with a regulator well before this and Bobken just passed the circuit details on to Peter having forgotten the source of the design.

You can use the circuit in post 1 for personal use. The output voltage can be set between 5 and 30 volts with the AD811AN package. If you bond a small heatsink to the chip you should be able to use it to up 36 volts. You will need to pay attention to voltage and power ratings of the components when setting up for the higher voltages in the range. The voltage reference, the 1K resistor from pin 2 to ground and the 470R resistor from pin 3 to ground will need re-sizing to suit different output voltages. The available output current will depend on the transfer characteristics and power rating of the shunt mosfet, the current source and the thermal management of the heat generated by these devices.

Any change in circuit parameters should be monitored on an oscilloscope to verify stability.

The post 1 design works well with line level analogue or digital circuits but it’s too noisy for vinyl preamps

Regards
Paul
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.